Verified65% confidenceFactTime unknown

目前顶尖模型在SWE-bench Verified子集上的解决率已突破50%

3

Sources

65%

Confidence

Medium-term (~90 days)

Relevance

5/31/2026

First Seen

Valid until: 8/29/2026

Sources

Amp编程代理：Sourcegraph推出的免费AI编码工具全面解析

rss

Related Entities

SWE-benchconcept

Related Claims

Unverified新一代模型在SWE-bench等真实软件工程任务上的通过率首次突破50%大关81% similar UnverifiedSWE-bench 发布初期顶级模型通过率不足 2%，当前领先系统通过率已突破 50%80% similar UnverifiedSWE-Bench上最强模型也仅能解决约50%的问题77% similar Unverified在 SWE-bench 上早期模型解决率不足 5%，2025 年主流产品已突破 50%77% similar Verified2024年顶级模型在SWEbench-Verified上的通过率约为50-70%76% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/8524

API

curl https://kongchang.com/api/v1/knowledge/claims/8524

MCP

get_claim(id=8524)